Shoulder Internal-Rotation Deficit Screening: A Side-to-Side GIRD Protocol for Overhead Athletes

A college pitcher walks into the training room in week two of fall ball and says his arm feels a little tight after a bullpen. He sits at 88 on the gun, his mechanics look clean on video, and the easy move is to file it under normal soreness and send him back out in four days. That is usually the exact moment a real screen gets skipped, and it is also the moment a side-to-side rotation check would tell you something the radar gun cannot.

Every throwing shoulder loses some internal rotation on the dominant side. That is not damage, it is years of humeral retrotorsion and posterior capsule adaptation doing exactly what they are supposed to do. The mistake is treating any internal-rotation loss as a red flag, or the opposite mistake, waving it off because he has always thrown like that. The number that actually separates a normal adaptive shoulder from an injury-risk pattern is not the internal-rotation deficit on its own. It is whether external rotation is compensating enough to keep total rotational motion equal between the two arms. This guide walks through the exact test position, the ranges that count as normal, and the calculation that tells you whether a shoulder needs a monitoring note or a modified throwing program.

Three things, no more: a table the athlete can lie flat on, a digital inclinometer, and a second pair of hands. A phone-based inclinometer app reads accurately enough for field use as long as you use the same device on both sides of the same athlete, and don't switch tools between a preseason baseline and an in-season re-test. Trying to stabilize the scapula and read the inclinometer yourself is exactly how a false 5 to 10 degrees creeps into the number, so get an assistant, or use a strap across the shoulder if you are working solo.

Skip the stabilization step and the reading isn't internal rotation at all, it's internal rotation plus scapular anterior tilt and forward translation, which can add ten to fifteen degrees of false range depending on how loose the athlete's scapular control happens to be that day. That single detail is the difference between a screening result you can trust across a season and one that just tracks how firmly you pressed on the shoulder blade this particular Tuesday.

The sequence takes about four minutes per athlete once you've run it a few times.

1. Test cold. Before throwing, lifting, or any shoulder warm-up — a warmed-up shoulder reads several degrees looser and hides the exact deficit you're trying to catch.
2. Start non-dominant. Position the athlete supine per the table above and measure the glove-side arm first, to reduce anticipatory guarding once you get to the arm that actually concerns you.
3. Rotate to end-feel. Passively rotate the forearm into internal rotation until you feel a firm capsular end-feel or your stabilizing hand feels the scapula begin to lift off the table. Stop exactly there and read the inclinometer.
4. Repeat for external rotation. Same setup, opposite direction, same stopping rule.
5. Run three trials per direction, per side. If the first trial differs from the next two by more than 5 degrees, discard it and average the remaining pair.
6. Repeat on the dominant, throwing side, then calculate the deficit before the athlete leaves the table so anything worth flagging gets a same-day follow-up, not a note buried in a spreadsheet.

Three trials is not overkill. A single unstabilized reading swings enough between examiners that two coaches measuring the same shoulder ten minutes apart can land 8 to 10 degrees apart. Add manual scapular stabilization and keep the same rater across an athlete's re-tests, and that spread tightens to somewhere around 3 to 5 degrees — tight enough to trust a week-to-week trend instead of chasing measurement noise.

Two numbers matter here, and neither one is the raw internal-rotation angle by itself. Glenohumeral internal-rotation deficit (GIRD) is the non-dominant arm's internal rotation minus the dominant arm's internal rotation. Total rotational motion (TROM) is internal rotation plus external rotation, calculated separately for each arm. TROM deficit is the percentage difference between the two arms: (non-dominant TROM minus dominant TROM) divided by non-dominant TROM.

Read that middle column carefully: a 15 to 25 degree internal-rotation loss on the throwing arm, with external rotation gaining back a similar amount, is the adaptive pattern seen in the large majority of healthy pitchers — not a problem to fix. What separates a normal shoulder from a flagged one is whether that trade happens in full. When external rotation fails to pick up the slack and TROM on the throwing side comes in more than 5 percent short of the glove side, that's the pattern tied to injury risk in the research below, regardless of what the raw GIRD number reads. Pair this mobility screen with a handheld dynamometry strength check on external rotation strength — a shoulder that's both ROM-restricted and weak in external rotation carries more risk than either finding alone.

Wilk and colleagues (2011) screened 296 professional pitchers preseason and followed them through the season for shoulder injuries serious enough to require time on the disabled list. GIRD by itself, measured as the raw side-to-side internal-rotation difference, did not separate the pitchers who got hurt from the ones who didn't. What did separate the two groups was total rotational motion: pitchers whose TROM deficit exceeded 5 percent compared to their non-throwing shoulder were roughly 2.6 times more likely to land on the disabled list that season. The finding reframes the whole screen — GIRD alone is close to noise, TROM deficit is the signal. The obvious limitation is the population: this is a preseason-to-season observational cohort in professional pitchers only, disabled-list placement undercounts the milder shoulder issues that never get reported, and pitch counts and workload weren't controlled for as a competing explanation.

Shanley and colleagues (2011) ran a similar screen-and-follow design in high school baseball and softball players — about 140 baseball athletes and 51 softball athletes tracked across a season by their athletic trainers. In the baseball group, a GIRD of 20 degrees or more was linked to roughly double the risk of a shoulder or elbow injury compared to athletes under that threshold. The pattern didn't hold in the softball players in the same study, which the authors attributed to the different arm path and lower internal-rotation demand of windmill pitching versus overhand throwing. That sport-specific split is a limitation worth sitting with: a GIRD cutoff validated in overhand throwers shouldn't be assumed to transfer directly to a windmill softball pitcher, a volleyball hitter, or a quarterback without separate validation, and the adolescent sample size here is small enough that a single injury or two shifts the risk ratio meaningfully.

Manske and colleagues' clinical review (2013) lands on the same rule: screen TROM deficit as the primary flag, treat isolated GIRD as background information, and reserve the term pathologic GIRD for a 20 degree or greater deficit paired with a TROM shortfall — not either marker alone.

Once you have both angles on both sides, the classification is mechanical, not a judgment call.

Re-screen on a fixed schedule, not just when someone complains: a preseason baseline, then every two to three weeks in season, and immediately after any shutdown longer than five days. An athlete flagged as pathologic usually doesn't need to be pulled from throwing outright — the more useful lever is reducing high-intent, max-effort throws while keeping low-intent volume up, alongside a structured posterior-capsule mobility block, then re-testing before adding intensity back. Clear an athlete back toward full throwing intent once TROM deficit comes back under 5 percent and GIRD is trending down, not once he simply says the shoulder feels better.

Most bad screening data doesn't come from a bad athlete, it comes from a rushed setup. The errors below account for the majority of readings that don't hold up on re-test.

• Scapular substitution. If you're not manually blocking the coracoid and anterior shoulder, the reading inflates apparent internal rotation, sometimes by 10 to 15 degrees, because the scapula tips forward instead of the humerus actually rotating.
• Elbow drift off 90 degrees. Even 10 to 15 degrees of extra elbow flexion or extension changes the rotational torque arm and shifts the reading independent of true shoulder mobility. Recheck elbow position between the internal- and external-rotation trials, not just at the start.
• Inconsistent end-feel. Firm end-feel needs a shared definition across raters — the moment resistance meaningfully increases, not the moment the athlete says it's uncomfortable. Mixing pain-limited stops with true capsular end-feel stops produces numbers that can't be compared session to session.
• Comparing to population averages instead of the athlete's own glove arm. A pitcher with 40 degrees of dominant-side internal rotation might be perfectly normal if his glove-side arm sits at 55 degrees, or seriously restricted if his glove side sits at 70. The athlete is always his own control.
• Testing right after practice. A shoulder tested immediately after a throwing session reads several degrees looser than a cold morning baseline. Standardize test timing — pre-practice is the default — and note the timing every time so a later comparison is apples to apples.